bhdump

Export browser history in JSON, CSV, and other formats.

bhdump is a cross-platform command-line tool that reads browser history databases and exports them as structured data. It supports all major browsers, provides powerful filtering via CEL expressions, and normalizes entries into a unified schema.

Features

• 11 browsers across three engine families (Chromium, Firefox, Safari)
• Auto-discovery of installed browsers and profiles on macOS, Linux, and Windows
• Output formats: JSON, JSONL/NDJSON, CSV, TSV
• CEL filter expressions for flexible querying (--where 'domain == "github.com"')
• Natural language dates (today, 7d, "last friday", "3 days ago")
• Noise filtering removes auth flows, tracking, CDNs, and other non-content URLs by default
• Safe reads by copying databases to a temp directory (avoids lock contention with running browsers)
• Usable as a library in addition to the CLI

Supported browsers

Installation

Pre-built binaries

Download a binary from the latest release.

Install script

curl --proto '=https' --tlsv1.2 -LsSf https://github.com/sagikazarmark/bhdump/releases/latest/download/bhdump-installer.sh | sh

Build from source

Requires Rust 1.85+.

cargo install --git https://github.com/sagikazarmark/bhdump

Usage

# Export all history as JSON (default)
bhdump

# Filter by time
bhdump today
bhdump 7d
bhdump "last friday"
bhdump --since 2024-01-01 --before 2024-07-01

# Filter by browser or profile
bhdump --browser chrome --browser firefox
bhdump --profile Default

# Choose output format
bhdump --format jsonl
bhdump --format csv
bhdump --format tsv

# Write to a file
bhdump --output history.json

# CEL filter expressions
bhdump --where 'url.contains("github.com")'
bhdump --where 'domain == "github.com" && visit_count > 5'
bhdump --where 'title.matches("(?i)rust")'

# Sort results
bhdump --sort -count       # most visited first
bhdump --sort url           # alphabetical by URL
bhdump --sort domain        # group by domain

# Other options
bhdump --limit 100          # cap number of entries
bhdump --dedup              # deduplicate by URL (keep most recent)
bhdump --visits             # individual visits instead of per-URL summary
bhdump --include-internal   # include chrome://, about:, etc.
bhdump --include-noise      # include auth, tracking, CDN domains

Subcommands

bhdump browsers             # list detected browsers and profiles
bhdump completions bash     # generate shell completions (bash, zsh, fish, etc.)
bhdump validate 'expr'      # validate a CEL expression without running a query

Filter expressions

The --where flag accepts CEL expressions evaluated against each history entry.

Available fields:

String methods: contains, startsWith, endsWith, matches (regex), size

Operators: ==, !=, <, <=, >, >=, &&, ||, !, in

Time shorthands

A positional argument serves as shorthand for --since:

Library usage

bhdump can also be used as a Rust library:

use bhdump::browsers;
use bhdump::filter::FilterConfig;
use bhdump::format::{OutputFormat, write_entries};

let sources = browsers::discover();
let (entries, errors) = browsers::read_all(&sources, None, None, false);

let filter = FilterConfig::default();
let filtered = filter.apply(entries).unwrap();

let mut stdout = std::io::stdout().lock();
write_entries(&mut stdout, &filtered, OutputFormat::Json).unwrap();

Comparison with similar tools

Several tools exist for extracting and querying browser history from the command line. The table below compares the ones that are most similar in purpose to bhdump.

When to use what

bhdump is designed for data export and pipeline use. If you need structured output with fine-grained filtering (e.g. piping history into jq, loading into a database, or feeding into analysis scripts), bhdump is the best fit. Its CEL expressions, noise filtering, and multiple output formats are built for this workflow.

browser-history is a good choice if you work in Python and want a lightweight library to integrate into your own scripts. It also supports bookmark extraction, which bhdump does not. However, it lacks filtering, noise removal, and only outputs basic JSON or CSV.

bhgrep is the right tool when you want to interactively search and browse your history from the terminal. Its fuzzy matching, TUI, and clipboard integration make it a quick "where did I see that page?" tool rather than a data export tool. It supports fewer browsers and has no structured export pipeline.

1History focuses on backing up browser history into a single SQLite database and visualizing it through a built-in web dashboard. If you want to maintain a persistent, deduplicated archive of your browsing history with visual analytics (top domains, daily page views), 1History is a good fit. It supports fewer browsers and output formats than bhdump, and has no filtering or pipeline features.

web-recap is a lightweight Go tool designed to extract browser history as JSON for feeding into LLMs. If your primary goal is piping history into an AI assistant for summarization or analysis, web-recap is purpose-built for that. It supports fewer browsers and output formats than bhdump, and has no filtering, sorting, or deduplication features.

Other tools

browser-gopher (Go) takes a different approach: it imports your history into its own SQLite database with a full-text index, then lets you search over the aggregated data. This is useful if you want a persistent, searchable archive of your browsing history over time, but it requires a separate populate step and is macOS-focused.

License

The project is licensed under the MIT License.